CN104915246A - High-configurable distributed real-time calculation engine based on workflow and control method - Google Patents
High-configurable distributed real-time calculation engine based on workflow and control method Download PDFInfo
- Publication number
- CN104915246A CN104915246A CN201410090455.7A CN201410090455A CN104915246A CN 104915246 A CN104915246 A CN 104915246A CN 201410090455 A CN201410090455 A CN 201410090455A CN 104915246 A CN104915246 A CN 104915246A
- Authority
- CN
- China
- Prior art keywords
- work
- working cell
- data
- unit
- treatment
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Abstract
The invention relates to a high-configurable distributed real-time calculation engine based on workflow and a control method to solve the problems that in the prior art, some data processing systems cannot well support high concurrency, real-time performance and high reliability. The engine comprises a workflow manager, a work unit monitor and a plurality of work units. The work units are connected to one communication adapter, and the workflow manager and the work unit monitor are connected to the communication adapter. According to the high-configurable distributed real-time calculation engine based on workflow and the control method, a workflow managing, operating and dispatching distributed calculation core is commonly composed of the workflow manager, the work unit monitor and the work units, the work units and work procedures on the work units can be organically organized, large-scale concurrence and real-time data processing can be well supported, and high availability is achieved.
Description
Technical field
The present invention relates to a kind of technical field of information processing, especially relate to and a kind ofly support that high concurrent, real-time, high reliability the height based on workflow can the real-time computing engines of partition cloth and control method.
Background technology
Along with the development of information industry and computer technology, the scale of emerging service rapidly increases, and as the seat reservation system of the Ministry of Railways, the form ordering system of Ali Taobao, process number every day with the request of hundred million notes, number of request even per second all may cross 1,000,000; And for example search engine, same needs the extensive request of response, and each request all needs to retrieve in TB DBMS the result that user wants.The and for example system such as electric power, traffic, communication, need high reliability, uninterruptedly provide service, service disconnection will bring huge loss.
During the now common Hadoop that increases income is distributed, mainly towards static data (as file, database), lack for stream data and support, each working link cannot perfect coordination work in addition, general mutual by intermediate data; And Hadoop installation and deployment more complicated; Development language is java, poor to the support of real-time.
Summary of the invention
The present invention mainly solves the problem that some data handling systems in prior art can not support high concurrent, real-time, high reliability very well, provides a kind ofly to support that high concurrent, real-time, high reliability the height based on workflow can the real-time computing engines of partition cloth.
Present invention also offers and a kind ofly support that high concurrent, real-time, high reliability the height based on workflow can the real-time computing engines control method of partition cloth.
Above-mentioned technical matters of the present invention is mainly solved by following technical proposals: a kind of height based on workflow can the real-time computing engines of partition cloth, comprise workflow manager, working cell monitor and some working cells, described working cell is connected on a communications adapter, and described workflow manager and working cell monitor are also connected on communications adapter;
Working cell: each working cell includes some progresses of work for the treatment of data, according to the form of process data, working cell arranges difference in functionality, by combination between each working cell, forms the data processing structure of process large-scale concurrent or real-time form;
Working cell monitor: the running status monitoring the progress of work on each working cell, and feed back to workflow manager;
Workflow manager: configuration, start-stop, scheduling and management work flow data, and according to working cell monitor feedack, the progress of work on working cell is dispatched.
The present invention constitutes the Distributed Calculation core of a Workflow Management, operation, scheduling jointly by workflow manager, working cell monitor and working cell, can workflow on each working cell of organic organization and working cell, make to support large-scale concurrent and real-time data process well, and there is high availability.In the present invention, workflow is the complete job flow process of certain business of process.Working cell is a link in work for the treatment of stream.Communications adapter to be responsible in various operating system the machine or across the communication between host processes.
As a kind of preferred version, when for process large-scale concurrent data structure, working cell comprises the first work for the treatment of unit of an input service unit and some process data, each first processing unit comprises the progress of work of process particular data, and input service unit is connected with each first processing unit respectively; Or working cell comprises the second work for the treatment of unit of an input service unit and process data, and input service unit and second unit of dealing with the work is connected, and unit of dealing with the work comprises multiple progress of work processing data respectively;
When for process real-time data structure, working cell comprise one for split task fractionation working cell, one for merging the 3rd work for the treatment of unit of the merging working cell of task, some Processing tasks data, these the 3rd work for the treatment of unit are connected in parallel on and split between working cell and difference working cell.Working cell can be divided into difference in functionality according to demand, or comprises the different progresses of work, carries out connection in series-parallel again and be combined into structure for the treatment of large-scale concurrent or real-time form data between each working cell.The progress of work that first working cell comprises process particular data represents that this working cell can only process a kind of data, multiple first working cell processes different pieces of information respectively, the division of labor processes, input service unit is when distributing data, each data are classified, distributes to the first corresponding work for the treatment of unit respectively.Second work for the treatment of unit comprises multiple progress of work processing data respectively, and these progresses of work are each can be processed data, and data are randomly assigned to each progress of work by the second work for the treatment of unit.3rd work for the treatment of unit comprises the progress of work of multiple process data, and these progresses of work are each can be processed data, and data are averagely allocated to each progress of work by the 3rd work for the treatment of unit.
As a kind of preferred version, also include operating system adapter and real-time database adapter, operating system adapter and real-time database adapter are connected to communications adapter.Operating system adapter is used for being connected with different operating system interface, and real-time database adapter is used for being connected with real-time database, carries out access visit.
Height based on workflow can a partition cloth real-time control method, comprises large-scale concurrent and real-time treatment step;
Large-scale concurrent treatment step comprises static treatment step and dynamic process step,
Static treatment step: data are distributed to input service unit by workflow manager, input service unit is classified to data, then every class data is distributed to the first work for the treatment of unit of such data of alignment processing;
Dynamic process step: data are distributed to input service unit by work manager, input service unit sends the data to one second work for the treatment of unit, and task is randomly assigned to some progresses of work and processes by work for the treatment of unit;
Real-time treatment step comprises: data are distributed to and split working cell by workflow manager, split working cell and Data Division is become some subdatas, be distributed to some 3rd work for the treatment of unit parallel processings, after all subdata process complete, then all send to merging treatment unit, by merging treatment unit, subdata is merged.Work stream data is undertaken rationally distributing in real time by working cell by the present invention according to demand, realizes the support to large-scale concurrent, real-time data process and high reliability.
As a kind of preferred version, also comprise the step that the progress of work of each working cell is controlled, its step comprises: working cell monitor monitors the information of each working cell, information comprises number of queues and processing speed, when certain working cell data increases, when processing speed is slack-off, Monitor Tag this working cell state in working cell is busy, and by this information reporting to workflow manager, workflow manager increases the progress of work of this working cell; When certain working cell request reduces, Monitor Tag this working cell state in working cell is idle, and by this information reporting to workflow manager, workflow manager reduces the progress of work of this working cell; The abnormality of monitor also simultaneously follow-up work process in addition, when certain progress of work occurs abnormal, Monitor Tag this working cell in working cell is abnormal, and by this information reporting to workflow manager, data are sent to the normal progress of work of state by workflow manager, and when calculating the required progress of work, always the progress of work of exception is rejected.This controls to make whole workflow be in stable, equilibrium state to the progress of work, ensure that the real-time of data processing, turn avoid the problem of task process failure.
Therefore, advantage of the present invention is: plurality of distribution can be carried out according to demand in working cell, each working cell can comprise the different progresses of work and be divided into difference in functionality, by forming the structure for the treatment of large-scale concurrent or real-time form data after combination between each working cell, large-scale concurrent and real-time data process can be supported well, and there is high availability.
Accompanying drawing explanation
Accompanying drawing 1 is a kind of structural frames view of the present invention;
Accompanying drawing 2 is a kind of syndeton schematic diagram of working cell when processing large-scale concurrent data in the present invention;
Accompanying drawing 3 is another kind of syndeton schematic diagram of working cell when processing large-scale concurrent data in the present invention;
Accompanying drawing 4 is a kind of syndeton schematic diagram of working cell when processing real-time data in the present invention.
Monitor 3-working cell, 1-workflow manager 2-working cell 4-communications adapter 5-operating system adapter 6-real-time database adapter 7-input service unit 8-first deal with the work unit 9-second deal with the work unit 10-split working cell 11-merge working cell 12-the 3rd deal with the work unit.
Embodiment
Below by embodiment, and by reference to the accompanying drawings, technical scheme of the present invention is described in further detail.
Embodiment:
A kind of height based on workflow of the present embodiment can the real-time computing engines of partition cloth, as shown in Figure 1, comprise workflow manager 1, working cell monitor 2, communications adapter 4, operating system adapter 5, real-time database adapter 6 and multiple working cell 3, each working cell includes again some each progresses of work.Working cell is connected on communications adapter, and workflow manager, working cell monitor, operating system adapter and real-time database adapter are all be linked on communications adapter.This working cell is according to the form of process data, and working cell can arrange different functions, and combines in communications adapter between each working cell, forms the data processing structure of process large-scale concurrent or real-time form.
As shown in Figure 2, when for process large-scale concurrent data structure, comprise static treatment Structure and dynamics process structure, static treatment arrangement works unit comprises the first work for the treatment of unit 8 of an input service unit 7 and some process data, here for three, each first processing unit comprises the progress of work of process particular data, and input service unit is connected with each first processing unit respectively.As shown in Figure 3, dynamic process arrangement works unit comprises the second work for the treatment of unit 9 of an input service unit 7 and process data, input service unit and second unit of dealing with the work is connected, and work for the treatment of unit comprises multiple progress of work processing data respectively.
When for process real-time data structure, working cell comprises one for splitting 10, one, the fractionation working cell of task for merging the 3rd work for the treatment of unit 12 of the merging working cell 11 of task, some Processing tasks data, and these the 3rd work for the treatment of unit are connected in parallel on and split between working cell and difference working cell.
Height based on workflow can partition cloth Run-time engine control method, comprises large-scale concurrent and real-time treatment step.Wherein large-scale concurrent treatment step comprises static treatment step and dynamic process step.
Static treatment step comprises: data are distributed to input service unit by workflow manager, and input service unit is classified to data, then every class data is distributed to the first work for the treatment of unit of such data of alignment processing.The method uses multiple working cell to provide service with load balancing mode, will be assigned on these working cells to this service data is on a rough average, and usual concurrent capability can increase by approximately linear.Its computing formula is as follows:
Formula 1:Ps=n × P × k
Wherein: P is the concurrent capability of single Service Instance; N is instance number; Ps is for being service cluster concurrent capability; K is distributed performance coefficient (0.9 <k < 1)
For Fig. 2, comprise an input service unit and three first work for the treatment of unit, each first work for the treatment of unit processes a kind of categorical data, tentation data A, B, C here respectively, and three first work for the treatment of unit process this three kinds of data respectively.During work, data are distributed to input service unit by workflow manager, data A is distributed to the first work for the treatment of unit of process data A by input service unit, and data B distributes to the first work for the treatment of unit of process data B, and data C distributes to the first work for the treatment of unit of process data C.
Dynamic process step comprises: data are distributed to input service unit by work manager, and input service unit sends the data to one second work for the treatment of unit, and task is randomly assigned to some progresses of work and processes by work for the treatment of unit.
The method uses multiple progress of work to provide service with load balancing mode, will be assigned in these progresses of work this service data stochastic averagina.
For Fig. 3, comprise an input service unit and one second work for the treatment of unit, the second work for the treatment of unit comprises three progresses of work, and three progresses of work are numbered 0,1,2 respectively, the following balanced way of data acquisition, as
I = route_data % instance_count
Wherein: I is example number, i.e. worker id; Route_data is route data, obtains from request msg, for recording id continuously; Instance_count is example sum.
If the data of input are 112, now the progress of work is 3, and result of calculation is 112%3=1, i.e. second progress of work, distributes to second progress of work and processes.Other input data process by that analogy.
Or can also other balanced ways be passed through, as
I = router_data / 100000
It is 100000 that this mode realizes each example capacity, according to period distribution request.This mode is generally used for scale, the expected scene of processing power, and the configuration of progress of work number is capacity scale, along with scale increases, revises this configuration, and ask can this growth of self-adaptation, can be properly routed to the corresponding progress of work.By the ability of dynamic dispatching example, 1,000,000 even more massive concurrent demands can be met.
Real-time treatment step comprises: data are distributed to and split working cell by workflow manager, split working cell and Data Division is become some subdatas, be distributed to some 3rd work for the treatment of unit parallel processings, after all subdata process complete, then all send to merging treatment unit, by merging treatment unit, subdata is merged.
The requirement of the most outstanding real-time is mainly reflected in data processing in enormous quantities and large-scale data process two kinds of scenes.Data processing in enormous quantities adopts large-scale concurrent treatment step above to ensure real-time.And large-scale data process, exemplify alarm completely in track traffic synthetic monitoring here and be described.
If a subway line there be n station, dispose n the 3rd working cell according to station number, will inquire about now alarm set on all stations.
Implementation procedure is as follows:
Fractionation working cell splits process, will ask " select * from alarm " be split as
select * from alarm where station_no=’Z01’;
select * from alarm where station_no=’Z02’;
...;
select * from alarm where station_no=’Zn’;
N subdata, and create affairs, then by three progress of work parallel processing of n sub-Data dissemination to correspondence; Each 3rd working cell completes inquiry and result is reported and be saved in corresponding affairs, then marks corresponding subdata treatment state, checks all subdata treatment states of affairs, when all subdata process complete, data is mail to and merges working cell; After merging working cell merges all results, delete affairs, recording processing result, and return Query Result.
Response speed calculates:
S = m + d + max(t
1,t
2,...,t
n) + r
Wherein: S is response speed; M is consuming time for splitting data; D is that route, transmission, affairs are consuming time; T1...tn is each subdata processing time; R is that amalgamation result is consuming time;
When the task processing time is much larger than fractionation, transmission, merging time (t
i>> m+d+r) time, be similar to:
S = t
max= max(t
1,t
2,...,t
n)
When split result is close to mean time, t
1=t
2=...=t
n=t
max, be similar to:
S = t
avg= ∑
1 nt
i/ n
Visible, by fractionation/merging treatment, the response speed of large task process can bring up to the processing time needed for the detachable minimum particle size of task.
A lot of system is as electric power, traffic, communication etc., and all need uninterruptedly to provide service, service disconnection will bring huge loss; Need to provide long-time high availability, so just relate to the step controlled the progress of work of each working cell, its step comprises:
Working cell monitor monitors the information of each working cell, information comprises number of queues and processing speed, when certain working cell data increases, when processing speed is slack-off, Monitor Tag this working cell state in working cell is busy, and by this information reporting to workflow manager, workflow manager increases the progress of work of this working cell; When certain working cell request reduces, Monitor Tag this working cell state in working cell is idle, and by this information reporting to workflow manager, workflow manager reduces the progress of work of this working cell; The abnormality of monitor also simultaneously follow-up work process in addition, when certain progress of work occurs abnormal, Monitor Tag this working cell in working cell is abnormal, and by this information reporting to workflow manager, data are sent to the normal progress of work of state by workflow manager, and when calculating the required progress of work, always the progress of work of exception is rejected.This ensure that whole workflow is in stable, equilibrium state, ensure that the real-time of data processing, turn avoid the problem of task process failure.Meet the requirement needing to provide long-time high availability.
Specific embodiment described herein is only to the explanation for example of the present invention's spirit.Those skilled in the art can make various amendment or supplement or adopt similar mode to substitute to described specific embodiment, but can't depart from spirit of the present invention or surmount the scope that appended claims defines.
Although more employ workflow manager, working cell monitor herein, make unit, the term such as communications adapter, operating system adapter, do not get rid of the possibility using other term.These terms are used to be only used to describe and explain essence of the present invention more easily; The restriction that they are construed to any one additional is all contrary with spirit of the present invention.
Claims (5)
1. the height based on workflow can the real-time computing engines of partition cloth, it is characterized in that: comprise workflow manager (1), working cell monitor (2) and some working cells (3), described working cell is connected on a communications adapter (4), and described workflow manager and working cell monitor are also connected on communications adapter;
Working cell: each working cell includes some progresses of work for the treatment of data, according to the form of process data, working cell arranges difference in functionality, by combination between each working cell, forms the data processing structure of process large-scale concurrent or real-time form;
Working cell monitor: the running status monitoring the progress of work on each working cell, and feed back to workflow manager;
Workflow manager: configuration, start-stop, scheduling and management work flow data, and according to working cell monitor feedack, the progress of work on working cell is dispatched.
2. a kind of height based on workflow according to claim 1 can the real-time computing engines of partition cloth, it is characterized in that:
When for process large-scale concurrent data structure, working cell comprises the first work for the treatment of unit (8) of an input service unit (7) and some process data, each first processing unit comprises the progress of work of process particular data, and input service unit is connected with each first processing unit respectively; Or working cell comprises the second work for the treatment of unit (9) of an input service unit (7) and process data, and input service unit and second unit of dealing with the work is connected, and unit of dealing with the work comprises multiple progress of work processing data respectively;
When for process real-time data structure, working cell comprise one for split task fractionation working cell (10), one for merging the 3rd work for the treatment of unit (12) of the merging working cell (11) of task, some Processing tasks data, these the 3rd work for the treatment of unit are connected in parallel on and split between working cell and difference working cell.
3. a kind of height based on workflow according to claim 1 and 2 can the real-time computing engines of partition cloth, it is characterized in that also including operating system adapter (5) and real-time database adapter (6), operating system adapter and real-time database adapter are connected to communications adapter (4).
4. the height based on workflow can a partition cloth Run-time engine control method, adopts the computing engines in claim 3, it is characterized in that: comprise large-scale concurrent and real-time treatment step;
Large-scale concurrent treatment step comprises static treatment step and dynamic process step,
Static treatment step: data are distributed to input service unit by workflow manager, input service unit is classified to data, then every class data is distributed to the first work for the treatment of unit of such data of alignment processing;
Dynamic process step: data are distributed to input service unit by work manager, input service unit sends the data to one second work for the treatment of unit, and task is randomly assigned to some progresses of work and processes by work for the treatment of unit;
Real-time treatment step comprises: data are distributed to and split working cell by workflow manager, split working cell and Data Division is become some subdatas, be distributed to some 3rd work for the treatment of unit parallel processings, after all subdata process complete, then all send to merging treatment unit, by merging treatment unit, subdata is merged.
5. a kind of height based on workflow according to claim 4 can partition cloth Run-time engine control method, it is characterized in that also comprising the step controlled the progress of work of each working cell, its step comprises: working cell monitor monitors the information of each working cell, information comprises number of queues and processing speed, when certain working cell data increases, when processing speed is slack-off, Monitor Tag this working cell state in working cell is busy, and by this information reporting to workflow manager, workflow manager increases the progress of work of this working cell; When certain working cell request reduces, Monitor Tag this working cell state in working cell is idle, and by this information reporting to workflow manager, workflow manager reduces the progress of work of this working cell; The abnormality of monitor also simultaneously follow-up work process in addition, when certain progress of work occurs abnormal, Monitor Tag this working cell in working cell is abnormal, and by this information reporting to workflow manager, data are sent to the normal progress of work of state by workflow manager, and when calculating the required progress of work, always the progress of work of exception is rejected.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410090455.7A CN104915246A (en) | 2014-03-12 | 2014-03-12 | High-configurable distributed real-time calculation engine based on workflow and control method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410090455.7A CN104915246A (en) | 2014-03-12 | 2014-03-12 | High-configurable distributed real-time calculation engine based on workflow and control method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104915246A true CN104915246A (en) | 2015-09-16 |
Family
ID=54084327
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410090455.7A Pending CN104915246A (en) | 2014-03-12 | 2014-03-12 | High-configurable distributed real-time calculation engine based on workflow and control method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104915246A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106326020A (en) * | 2016-08-24 | 2017-01-11 | 浙江浙大中控信息技术有限公司 | Distributed communication system and distributed communication method applied to static cluster |
| CN108427603A (en) * | 2018-01-10 | 2018-08-21 | 链家网（北京）科技有限公司 | A kind of task allocating method method and device |
| CN108920268A (en) * | 2018-07-12 | 2018-11-30 | 湖南商学院 | A kind of distributed system workflow processing method and workflow engine system |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102137125A (en) * | 2010-01-26 | 2011-07-27 | 复旦大学 | Method for processing cross task data in distributive network system |
| CN102236578A (en) * | 2010-05-07 | 2011-11-09 | 微软公司 | Distributed workflow execution |
| CN102592198A (en) * | 2011-12-30 | 2012-07-18 | 福建富士通信息软件有限公司 | Workflow engine supporting combined service |
| CN103279390A (en) * | 2012-08-21 | 2013-09-04 | 中国科学院信息工程研究所 | Parallel processing system for small operation optimizing |
| CN103414761A (en) * | 2013-07-23 | 2013-11-27 | 北京工业大学 | Mobile terminal cloud resource scheduling method based on Hadoop framework |
-
2014
- 2014-03-12 CN CN201410090455.7A patent/CN104915246A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102137125A (en) * | 2010-01-26 | 2011-07-27 | 复旦大学 | Method for processing cross task data in distributive network system |
| CN102236578A (en) * | 2010-05-07 | 2011-11-09 | 微软公司 | Distributed workflow execution |
| CN102592198A (en) * | 2011-12-30 | 2012-07-18 | 福建富士通信息软件有限公司 | Workflow engine supporting combined service |
| CN103279390A (en) * | 2012-08-21 | 2013-09-04 | 中国科学院信息工程研究所 | Parallel processing system for small operation optimizing |
| CN103414761A (en) * | 2013-07-23 | 2013-11-27 | 北京工业大学 | Mobile terminal cloud resource scheduling method based on Hadoop framework |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106326020A (en) * | 2016-08-24 | 2017-01-11 | 浙江浙大中控信息技术有限公司 | Distributed communication system and distributed communication method applied to static cluster |
| CN106326020B (en) * | 2016-08-24 | 2019-06-18 | 浙江浙大中控信息技术有限公司 | Applied to the distributed communication system and method on static cluster |
| CN108427603A (en) * | 2018-01-10 | 2018-08-21 | 链家网（北京）科技有限公司 | A kind of task allocating method method and device |
| CN108920268A (en) * | 2018-07-12 | 2018-11-30 | 湖南商学院 | A kind of distributed system workflow processing method and workflow engine system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108335075B (en) | Logistics big data oriented processing system and method | |
| US20180025057A1 (en) | M x n dispatching in large scale distributed system | |
| US10521396B2 (en) | Placement policy | |
| CN102043682B (en) | Workflow exception handing method and system | |
| CN106817408B (en) | Distributed server cluster scheduling method and device | |
| US7650347B2 (en) | System and method for job scheduling and distributing job scheduling | |
| CN104601664B (en) | A kind of control system of cloud computing platform resource management and scheduling virtual machine | |
| CN105631026A (en) | Security data analysis system | |
| CN104021194A (en) | Mixed type processing system and method oriented to industry big data diversity application | |
| CN104969213A (en) | Data stream splitting for low-latency data access | |
| US20060095914A1 (en) | System and method for job scheduling | |
| CN110865997A (en) | Online identification method for hidden danger of power system equipment and application platform thereof | |
| US9747130B2 (en) | Managing nodes in a high-performance computing system using a node registrar | |
| US20210392198A1 (en) | Data Center Management System | |
| CN105071994B (en) | A kind of mass data monitoring system | |
| CN106131227A (en) | Balancing method of loads, meta data server system and load balance system | |
| CN106528853A (en) | Data interaction management device and cross-database data interaction processing device and method | |
| CN104915246A (en) | High-configurable distributed real-time calculation engine based on workflow and control method | |
| CN103473848A (en) | Network invoice checking frame and method based on high concurrency | |
| CN103270520A (en) | Importance class based data management | |
| CN102479354A (en) | Data processing method and system based on workflow | |
| CN107423336A (en) | A kind of data processing method, device and computer-readable storage medium | |
| Zhou et al. | A multi-agent distributed data mining model based on algorithm analysis and task prediction | |
| CN116777182A (en) | Task dispatch method for semiconductor wafer manufacturing | |
| CN110390466A (en) | A kind of multidimensional visualized O&M managing device based on cloud SOA framework |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150916 |
|
| RJ01 | Rejection of invention patent application after publication |